Iqiso Eliphazamisayo Le Imithombo Engaziwa Kanjani Kuko Emakilasini Esimanje
Thina njengothisha sonke sesike sabhekana nesikhathi esingakhululekile lapho sigcina sesigqiba ukuhlola iqoqo lezinkulumo ebusuku. Uyabheka umbhalo womfundi—mhlawumbe ofunda i-English Language Learner—that uqukethe amagama acebile kakhulu kanye namaphethini emisho ayinkimbinkimbi angahambisani nhlobo nomsebenzi wakhe wangaphambilini ekilasini. Ukusola okusheshayo kuwukukhohlisa ezifundweni, kodwa uma ususa umbhalo ngohlelo olujwayelekile lokuthola ukufana (similarity checkers), akukho okukhangiswayo. Kuthiwani uma abafundi bakho bekopisha emithonjeni ongakwazi ukuyifunda, bebeka izindatshana ezisuka kolunye ulimi ngqo e-English? Lesi simo, esihlanganiswe nokukhula okusheshayo kwe-generative AI, sesishiye othisha abaningi bezizwa bethukuthele futhi bengenamandla. Ukuncika ku-AI detection kuphela akusanele ukugcina ubuqotho kwezemfundo emabangeni ethu asehlukahlukene kakhulu futhi athuthukile ngobuchwepheshe.
Izithiyo Zolimi kanye Nokutholwa Kwe-AI Okungahambi Kahle
Ikilasi lesimanje indawo ehlukahlukene nelivulekile ngokolimi, eletha ingcebo enkulu yamasiko kanye nezinselelo ezithile mayelana nokuthi umsebenzi womfundi uvela kuye yini. Uma abafundi behlangabezana nezithiyo zolimi ekukopeni (plagiarism), bangase baphendukele ekuguquleleni izindatshana zamanye amazwe ezingaziwa kalula, ngaleyo ndlela baphazamise ama-conventional similarity checkers angabheki kuphela ezizindeni zase-English. Ngaphezu kwalokho, ukungena kwe-generative AI emisebenzini yabafundi kushintshe ngokuphelele isimo sokungathembeki ezifundweni. Sibe sesibhekene nosongo oluyinkimbinkimbi olunezimpande ezimbili: i-plagiarism eguquliwe (translated plagiarism) nombhalo okhiqizwe umshini ngokunemba.
Kubalulekile ukuba siqonde imikhawulo yobuchwepheshe yamathuluzi okuthola i-AI akhona. Lawa masistimu asebenza ngamathuba okwe-statistics, ehlaziya amamethrikhi afana ne-perplexity ne-burstiness ukuze aqagele ukuthi umbhalo ubhalwe ngumuntu noma umshini. Ngenxa yokuthi asebenza ngamathuba (probabilistic), kuba nezikhubekiso ezinkulu, ikakhulukazi ama-false positives kanye nama-false negatives. I-false positive—lapho umbhalo womfundi wangempela ubikwa ngokungathi uvela ku-AI—ingalimaza ngokungenakulungiseka ubudlelwano phakathi kukathisha nomfundi futhi ibangele ukukhathazeka okukhulu kumfundi. Ngokuphambene, i-false negatives ivumela ukungathembeki okusezingeni eliphambili ezifundweni kudlule phakathi kwemifantu. Njengothisha, kufanele sivume ukuthi amathuluzi okuthola awawona ama-arbiters angapheli eqiniso. Angamathuluzi angaphelele angakwazi ukuqhuba esikhundleni sokuqonda okujulile okuba nothisha ngamakhono omfundi wakhe nokukhula kwakhe.
Ukushintsha Kokufundisa Okusekwe Ekulinganisweni Kwendlela (Process-Based Assessment) Futhi Nokuqonda Okusebenzayo (Authentic Learning)
Phambili, kufanele sishintshe ukugxila ekutholeni okungabheki—siye ekuthatheni izixazululo zokufundisa (pedagogical solutions) ezisebenzayo. Impendulo kulezi zinselelo eziyinkimbinkimbi ilele ku-process-based assessment kunokuthembela ngokuphelele emiphumeleni yokugcina. Ngokugqamisa indlela yokubhala, singakha ukuzethemba komfundi futhi siqinisekise ukuthi ukufunda okusebenzayo kwenzeka ngaphandle kokuhlala kulawulwa ngokuqondisa njalo ama-algorithms angamaphutha.
Isu lokuqala ukusebenzisa umlando wenguquko (document version history) njengengxenye ejwayelekile yenqubo yokuhlola. Amapulatifomu afana ne-Google Docs avumela othisha ukuthi babuyekeze inqubo yokudweba (drafting) yonke, bebheka ukuthi umfundi wakha kanjani izimpikiswano zakhe ngokuhamba kwesikhathi. Ukuphazima kokuvela kwezigaba ezinkulu zombhalo ongenaphutha ngaphandle komlando wokubhala wangaphambilini kuwuphawu oluqinile lokuthi kungaba i-translated plagiarism noma ukwakhiwa nge-AI. Lo mkhuba ushintshela ingxoxo ekusoleni uye ekuxoxisaneni ngokubambisana mayelana nenqubo yokubhala uqobo.
Isu lesibili lihilela ukwenza abafundi babhale ngokulandelana (iterative drafting) nokuhlola okuqhubekayo okuqondisa (continuous formative assessment). Uma imisebenzi ihlukaniswa ibe yizigaba ezitholakala kalula—njengokucabanga (brainstorming), ukuhlelela (outlining), ukubhala (drafting), nokubuyekeza (revising)—abafundi mancane amathuba okuthi baphazamiseke bese beya ekungathembekini ezifundweni. Ukunikeza impendulo (feedback) esigabeni ngasinye kwakha indawo yokusekelwa (scaffolded environment) lapho uthisha azi ngokujulile ukuthi imibono yomfundi ithuthuka kanjani. Le ndlela ivimbela ngokwemvelo ukusetshenziswa kwemithombo yangaphandle engakaziwa noma amathuluzi e-AI, ngoba umfundi kufanele abonise njalo ukuqonda kwakhe okukhulayo.
Isu lesithathu ukwakha imiyalezo (prompts) ecacile kakhulu, incike kumongo (context-dependent). Izihloko ezijwayelekile zezindatshana kulula ukuzikhomba ku-generative AI noma zitholakale ezindabeni zangaphandle ezazivele zikhona. Kunalokho, kufanele sakhe imisebenzi edinga ukuthi abafundi baxhumanise imiqondo yekhosi nolwazi lwabo siqu, izingxoxo zakamuva ekilasini, noma izehlakalo zendawo ezithile kakhulu. Idizayini yomsebenzi eyiqiniso (authentic assignment design) iphoqa abafundi ukuba bahlanganyele ngokujulile nendaba, okwenza kube nzima kakhulu kubo ukweqa umsebenzi wokuqonda odingekayo ukuze bakhiqize impendulo yabo yoqobo.
Ukulungisela Ikusasa Ngokuzethemba kanye Ne-Professional Expertise
Isimo semfundo siyashintsha ngokungangabazeki, futhi izinselelo ze-translated plagiarism kanye ne-generative AI zihlale zikhona. Nakuba umuzwa ungaba ukufuna ithuluzi elihle lokuthola i-AI, singavikela ubuqotho ngokuhlelwa okuphelele okuhlanganisa ubuchwepheshe nemfundiso (pedagogy). Ngokwamukela i-process-based assessment, ukwakha imisebenzi eyiqiniso, nokugcina ukugxila ekukhuleni komfundi, singaqinisekisa ukuthi amakilasi ethu ahlala eyizindawo zokufunda okungempela. Njengothisha, ithuluzi lethu elikhulu akusilo i-algorithm, kodwa i-professional expertise yethu kanye nokuzibophezela kwethu ekukhuliseni ubuqiniso bomfundi. Sinamandla okuzivumelanisa, ukuqondisa abafundi bethu, futhi siphumelele kule nkathi entsha yemfundo.
